But you won’t retain your customer base, because your competitors will invest in the new innovation and undercut your price. You will lose customers and you will start losing money.
…the competitors *might *invest in the new innovation. They might not. And they might try and undercut my price. Or they might instead choose to hold their prices and take a larger profit margin instead. And if they do undercut my price: then I’ll do what businesses have always done, find different ways to innovate, offer superior customer service or add value to the services that I offer. I may or may not succeed. My photography business is still going strong despite the fact that people undercut me and my primary camera is six years old. So I think the odds would probably be in my favor.
There is no rule that decrees that if I don’t “adopt an innovation I begin losing money every quarter until bankruptcy.” That just isn’t how it works.
You’re using the wrong analogy. Compare to tractors. Compare how much farming a single farmer can do with and without a tractor. Realize that the commodity price for produce is a result of a market price contributed to by farmers that also have access to tractors.
Unless you can sell your hand raised crops outside the normal market (which is not a valid comparison), you will lose money every quarter until bankruptcy raising crops by using horse drawn plowers instead of a tractor. See The Grapes of Wrath or historical documents on that era for what happened to the farmers who couldn’t afford tractors.
In any free market with a reasonable amount of participants, the competitors will adopt innovations that lead to large, clear cut, and immediate reductions in cost, which is what autonomous trucks and delivery vehicles will do.
…my “analogy” fit all the criteria to fit your definition of “forced demand.” It wasn’t “wrong.” Its how business works. You’ve shifted your definition. The analogy you have presented is akin to asking how much photography can a photographer do without a camera.
What you are really talking about is “disruptive innovation”. It would be helpful if you could use real world economic terms because they have established definitions and then I can understand what you are talking about. Your concept of “forced demand” doesn’t make sense without the additional context provided by the definition of “disruptive innovation.”
We don’t know what autonomous trucks and delivery vehicles will do yet. The technology just isn’t there yet and we won’t know if they will “lead to large, clear cut, and immediate reductions in cost” or if they will be too expensive for anyone to actually adopt. They sure have the potential to disrupt the industry, I don’t deny that. But what will actually happen? And to what extent? I think we will have to wait a few years (or decades) before we can find out.
Do you have an actual…basis for believing this? That is, you have posited a scenario that you “doubt” they will be cheaper. And you don’t trust other’s methodology and I have a hunch you didn’t look up the present cost of the sensors, the cost of a marginal copy of software (hint…it’s really, really, really small…), the present cost of the chips, and the expected future cost of the electronics in the next 3-5 years assuming a similar trend to the last 100,000 electronic product releases over the last 30 years.
Average annual salary for a truck driver is $40,000. So exactly what did you think goes into an autonomous car…? Did you think these cars were going to be sold like tractors? That isn’t the plan…they won’t be sold, it’s infeasible…
…believing what? That We don’t know what autonomous trucks and delivery vehicles will do yet? The technology just isn’t there yet and we won’t know if they will “lead to large, clear cut, and immediate reductions in cost” or if they will be too expensive for anyone to actually adopt?
Yep. I believe that.
The word “doubt” was not used in my post. You highlighted what you consider to be “doubt” but you completely ignored the bit that was immediately before it. I make no judgement at this time whether or not things will be cheaper. And that should be bleeding obvious.
Strawman. I doubted the veracity of a couple of cites. I’m sure you doubt the methodology of stuff all the time. That doesn’t mean I “don’t trust” every methodology.
I don’t need to know the cost. Because…
…the autonomous truck that is going to replace the truck driver doesn’t exist yet. I’m not investing in autonomous trucks. I don’t need to know the present cost of the sensors, the cost of a marginal copy of software , the present cost of the chips, and the expected future cost of the electronics in the next 3-5 years assuming a similar trend to the last 100,000 electronic product releases over the last 30 years. I don’t need to know any of this.
When the technology is ready to hit the market *then *I’ll take a look at what’s available and decide 1) does my business need this technology and 2) whether or not my business can afford this technology. But until then I don’t really give a fuck.
Ok, we’re conflating two issues. You are talking about whether to invest your real money in a specific company. I am trying to defend a specific technology.
If we rolled back time to 1998 or so, I’d be here explaining that the advantages of the internet are huge. There’s real genuine value in having a worldwide network of computers. And advantages to business.
You might rightfully ask which companies are going to make money, and how much, and when. And I guess piss me off by figuring the internet is just a fad and you don’t know if anyone will be able to make any money from it at all, I mean you have all these doubts, and computers are expensive, and presently demonstrated hardware and network speeds are really slow…
Well, I don’t know what I would have said, but here we are 22 years later, and honestly any advice I gave then would be at least partly dead wrong. Like 10 companies have become megacorps (ebay, Amazon, Google, Facebook…) and the rest have all folded. The primary ways money is being made is not by selling information on the internet, most of the sites are still free, but by advertising and selling portable access devices that among other things connect to the internet.
So as for which companies to invest your pennies : look, here’s the best analysis I know. I actually don’t think GM/Ford, both companies with massive billion dollar autonomy efforts, have “got this”. I think the nature of the technology - this is knowledge from the periphery, looking at the state of the art in machine learning - is going to lead to a small number of big winners who are specialized *software *companies. And it won’t be from selling autonomous vehicle services at all, it’ll be from selling autonomy itself to a far larger pool of clients.
But that’s just what I think, and while this might mean your best move is to do “long, Alphabet Inc.”, it could easily be a spin-off company from one of the others that ends up winning the platform wars for this “cloud based autonomy framework*” and taking the whole pot.
Sort of like the early 70’s when Datsun (Nissan now), Honda and Toyota started to make cheep high mileage cars and Ford and Chevy sat on their thumbs.
Now, I am one of he skeptics of SDC’s. I very, very much doubt it will ever be an option for myself. I still don’t enjoy a high mileage car, because it’s not an option either. Non exist that will do what I need, and I very much doubt ever will. SDC’s will however change driving as we know it, and will help many.
Although I hate driving and would be happy to have a car that can drive for me, I worry about how they’d preform in New England winters. I kind of worry that we’ll end up with cars that are self-driving…except need us to take over in the winter. People are not going to drive better if they’re only doing it in bad weather, and they drive terribly in bad weather as it is (case in point: we got less than 2" of snow on Thursday. No one plowed until mid-morning, and my 50 minute commute before they plowed took 90 minutes because three people racing around like morons got into separate crashes. Had no one been reckless, they likely would’ve been wreckless too because the roads were just snowy, not especially icy). Hopefully, some of the new technology that they’re testing to help with bad weather driving in general like this, will both work and be put into self-driving cars.
…I’m not conflating anything. I’m directly responding to your questions. You seem to be interested in having a conversation I’m not interested in having. If this was twitter I’d say “tag me out from this conversation.” As its not twitter, I’ll just say I will not be responding to your posts directed at me from this point on.
It might do better than people think. (Direct link to the video, it’s only 30 seconds.) It is a Tesla Model 3 AWD using autopilot on an icy highway. The car skidded, slid sideways across a lane, and then the autopilot recovered and got the car going straight again. The driver claims that only autopilot (and traction control) were used, and he didn’t provide any input to recover from the skid.
I know the highly experienced winter drivers on this forum would have kept the car in its original lane, and not let the skid go that far. I know that many of the winter drivers I see on the roads would have wound up stuck in the snow on the shoulder, or stopped in traffic facing the wrong way.
This is my analysis of that video. The car was probably going too fast for conditions (almost by definition), or at least it was going faster than that car in front of it. That isn’t the autopilot’s fault, as the speed is set by the driver. The driver claims black ice, and that might be true, but it also looks like wet slush covering the roads. That can be as bad as ice, because it is very easy to hydroplane in thick slush. This also shows that all wheel drive isn’t a magical charm against skidding. Nobody here claims that it is, but I think a lot of people get deceived into believing it is, because all wheel drive works so well getting the car to go.
I’ve used autopilot in the rain, and I increased the following distance to what I felt was appropriate. I had the opportunity to use autopilot in the snow last week, but I didn’t do it. It’s not so much that I think I can recover from a skid better than it, but rather that I think if I’m driving I’m going to have a better feel for how slick the road is, and adjust my speed accordingly.
Handling these type of conditions is something that fully autonomous cars will have to do. Even more than handling a skid, the systems will have to know what to do to avoid skidding in the first place. They might eventually be very good at that, because autonomous systems should be know how much traction is available. They might even be able to make decisions such as, “other vehicles report conditions are too poor, so I refuse to make this trip.” That will just reinforce Enipla’s argument of why he’ll never have a self driving car, but sometimes staying off the roads is the best option.
I also think vehicles that are fully autonomous only in some conditions are fine. I don’t want to give up the limited autopilot I have now, just because it might not be appropriate to use on a few snow days per year.
GPS seems to work well even in bad weather - at least, my phone navigation seems to do the job. Maybe it’s supplemented by cell signals then. Are there any conditions in which the GPS component would not be expected to work well at all?
And re the cars working in all conditions: the car might need to have some detector that says “it’s too snowy / too much standing water / too many meteorites falling, you have to switch to manual”… and it would need to do so in a way that allows the human driver to respond in time, or it would need to be able to bail to the side of the road if the driver did not respond. And it would need to have controls so that if the human driver reacted badly (as you might do, if you had been snoozing or something), the driver couldn’t kill himself by accident. *
The need to have a human driver available is one reason why I don’t think they’d be a good choice for taking Junior to Little League and similar, ditto older but incapable passengers. The “safely handle any situation that might arise” ability may be a bit longer in coming than “can derp along in traffic in nearly every scenario”.
We were at a mall with a Tesla store, a few weeks ago. Because we are in the market for a new car, we sat in one and asked lots of questions. It has an intelligent cruise control feature, where it won’t approach a car in front too closely even if you have the cruise control set higher. We had that in a Toyota we rented last year and I liked it. However, it supposedly also had a lane-change feature. Not one to warn if you drifted across the line… one that would change lanes for you.
Pretty sure that would freak me the hell out.
In any case, it’s a nonissue, as the price is about 20K higher than whatever we might choose to purchase.
- On purpose, now… sets up some interesting legal puzzles. What if you murder someone while driving somewhere, and somehow exit the car safely while it’s moving, and the car is set to drive somewhere outside the current jurisdiction? Who arrests you? Where are you tried? Would states pass extradition rules that make it an automatic “yes” when such a request is made? I assume suicide doesn’t pose the same legal issues, unless place of death is important for some reason.
It does. Larry Magid, tech columnist for the Mercury News, has a new Tesla and has written about the lane change feature. It can also take exit ramps from one freeway to another without manual intervention. He’s done it.
It took over 3 seconds to recover, and that’s a hell of a long time. Hopefully it’ll improve significantly soon.
GPS signals are not accurate enough, IMO for self driving cars because they can put you to within a couple of feet, not a couple of inches. This is an image of my track from a run (race) earlier this week. I can assure you that at no time was I in or floating over the River. On the way out (up), I was in the right side of the SB outer lane, about 2’ from the edge of the bridge. On the way back (down), I was on the left side of the inner lane (the SB lanes were closed to traffic). I can also assure you that I was on paved road the whole time, unlike the view at the bottom & the right where I’m going thru trees.
GPS needs a clear view of the sky; it doesn’t work well in urban canyons, tunnels, or heavily forested areas. As for the comment upthread that the SDC will use GPS to get into the ballpark & then use visual clues (firehydrant, mailbox, etc.) for the rest I’d suggest trying to use Google street view for your local strip mall. I’ve found it doesn’t have all of the roads near me let alone the paved non-roads, & I live in Greater Suburbia, not some remote, rural area.
Will we get there, probably eventually; are we there today, no way!
And it was going straight down the road when it lost it.
I’m assuming the “visual clues” means something more precise than just Google Street View. Instead, I think the computer will use a 3D map of the area, built up over time from multiple passes by multiple vehicles, of all objects in the area. I don’t expect the on-board computer to store all of this data locally. Instead, it will constantly communicate with a larger database. In fact self-driving cars are one of the proposed applications for the 5G networks now being developed.
What I was saying about Street View is that there are plenty of places that that hasn’t even been to yet & they’ve been at it a while. You’re going to need something that maps every road & parking lot that is more precise that that is.
I thought the 5G was to communicate to other cars & traffic control devices (the new version of a traffic light); if it’s used for general navigation then that’s an issue. 5G is faster but shorter range than 4G, which means even more antennas; yet there are significant (rural) areas that don’t have 4G coverage today. What about when a significant weather event (ice storm or hurricane) knocks out power to an area for a week or two? Are all of these new antenna sites have long-term batteries to communicate or will cars just not be able to drive because they’ve lost their ability to communicate to a central server? Who’s paying for all of this new infrastructure?
While I wouldn’t mind a self-drving car, I certainly want other people to be in self-driving cars. I’ve seen [the collective] them drive, and they’re terrifying at it.
Imagine if self-driving cars killed twenty-seven thousand people every year. Horrific? Well, since we’ve lost 37,000 souls per year in 2016 and 2017 to human-driven cars, that would represent saving ten thousand lives every year.
Can’t come soon enough.
Yeah, it isn’t clear what caused the skid in the first place. If it really was a situation of the car driving straight with no turning, acceleration, or deceleration, then I think the skid would have happened under autopilot or with the human driver in the same conditions. The same if it was just a very minor control input; a human driver could have easily done the same thing.
These are my guesses based on my experience with the same autopilot. Perhaps it was applying power in order to maintain speed through the slush. Perhaps it was braking to maintain speed on a downhill portion or because it thinks the car in front is coming into its lane. Perhaps it is generating steering inputs to stay in the lane, or to adjust itself between the lines.
The slush on the road may be confusing the lane tracking. During a heavy rain while driving into the sun, even with polarized glasses, the glare on the road was so bad that I had a very hard time seeing the lane lines, but the car did it perfectly. I was very impressed. However, with snow and slush obscuring the lines, and tracks cut through the snow, it may have gotten confused. I know mine occasionally gets confused in certain spots where the lane markings have been changed.
As referenced in my first post about this, I think the ultimate cause was going too fast for the conditions. That is something completely in control of the driver, though the autopilot will slow down for obstructions and some curves.